Latest Headlines from Nourish | The Nourish Mission

OpenAI says China‑linked campaigns used ChatGPT

What's happened

OpenAI has published research saying it has identified two small, likely China‑linked influence operations that used ChatGPT to generate social media posts and political cartoons about US data centres and tariffs. The company has said the campaigns gained little authentic engagement and found no evidence they meaningfully shifted public debate.

What's behind the headline?

What happened and why it matters

OpenAI has published a report saying it has identified two influence operations that used its ChatGPT model to scale posts, comments and cartoons about US data‑centre development and tariffs. One operation, which OpenAI calls "Data Center Bandwagon," produced comics and comments blaming data centres for higher electricity bills. The other, "Tech and Tariffs," produced cartoons critical of US tariff and tech policy.

What the evidence shows

  • OpenAI says the campaigns were small, short lived and gained little authentic engagement. It says one cluster likely operated from China and used ChatGPT prompts in Simplified Chinese that requested English outputs.
  • OpenAI reports the operators repeatedly asked that images not include Xi Jinping, suggesting deliberate messaging control.
  • The company says it found no indication the campaigns changed the volume or tone of public debate about data centres.

What drives this story now

  • Local opposition to data centres has increased across the US, producing stalled projects and proposed federal moratoria such as the bill from Senators Sanders and Rep. Ocasio‑Cortez. That domestic debate creates ready material for influence operators to amplify.
  • Republican lawmakers and industry figures are already citing foreign influence as an explanation for anti‑data‑centre sentiment; OpenAI's report will feed that political narrative and push federal agencies to probe foreign meddling.

Consequences and likely next steps

  • The report will increase pressure on regulators and lawmakers to investigate foreign influence in tech policy. Congress and executive agencies will likely request briefings and demand more transparency from AI firms about misuse of their models.
  • Firms that operate AI models will tighten detection and take‑down processes and will likely publish more attribution reports. That will raise debates about platform responsibility and transparency.
  • The finding will harden partisan lines: pro‑AI actors will use it to argue opponents are being amplified by foreign actors, while critics will continue to point to local environmental and infrastructure harms as independent reasons to oppose data centres.

Bottom line

This episode shows influence operators are testing AI tools to amplify existing US flashpoints. The campaigns did not succeed at scale, but they will force political and regulatory actors to treat AI‑enabled influence as part of broader tech and national‑security debates.

How we got here

Data‑centre construction has become a contested political issue in the US because communities worry about energy and water use. OpenAI and other firms say data centres are critical to AI capacity; lawmakers and activists have argued for moratoria or tighter oversight amid rising local opposition.

Our analysis

OpenAI has said it uncovered two operations that used ChatGPT to craft posts and cartoons targeting US debates over data centres and tariffs; Ben Nimmo, principal investigator on OpenAI's intelligence team, told reporters the groups prompted ChatGPT in Simplified Chinese and asked for English outputs (Business Insider, Politico, Axios). OpenAI said the "Data Center Bandwagon" posts blamed data centres for rising electricity bills and created comic strips showing a businessman and shocked family; it said the activity "did not have meaningful influence" (Al Jazeera, Axios). Business Insider reported OpenAI found one cluster likely tied to a private Chinese tech firm working for provincial clients and noted the operators instructed the model not to depict Xi Jinping. Axios flagged that the campaigns latched onto an existing debate rather than creating it and quoted Nimmo: "This was not a case of an influence operation creating a debate." Politico and Business Insider noted the report will amplify Republican claims that foreign actors are behind anti‑AI sentiment, while the New York Post and allied commentators have already urged federal probes and cited reports from groups such as the Bitcoin Policy Institute alleging foreign funding in anti‑data‑centre campaigns. Several local reporting outlets have detailed strong community pushback to specific projects — from Coachella to Nottingham to Box Elder, Utah — showing why influence operators find receptive narratives. Taken together, the sources show agreement that AI tools are being tested by foreign operators, disagreement over the scale and effectiveness of the campaigns, and sharp political use of the findings by opponents and defenders of data‑centre expansion.

Go deeper

  • Will federal agencies open a formal investigation into these alleged China‑linked campaigns?
  • How will AI companies change their monitoring and attribution of misuse of their models?
  • Will evidence linking specific Chinese entities to influence operations emerge publicly?

More on these topics

Latest Headlines from Nourish | The Nourish Mission